Copeia,1997(4), pp. 690-696

Ecology of Cichla (Cichlidae) in Two Blackwater Rivers of Southern Venezuela

KIRK o. WINEMILLER, DONALD C. TAPHORN, AND MIELLO BARBARINO-DuQUE

We investigated the ecology of peacock (Cichla spp.), diurnal piscovores and important gamefishes, in the Rio Cinaruco (Rio Orinoco drainage of the south- ern Venezuelan llanos) during the dry seasons of 1992-1993 and the Rio Pasimoni (Rio Negro-Amazonas drainage) during early 1993. In the Cinaruco, three Cichla species partitioned available habitats: C. intermedin near structure in primary river channel within or near swift current, C. orinocensisin shallow water along the shore- line of lagoons or slow channel reaches, and C. temensisin deeper waters along sandbanks and steeper rocky banks of lagoons and the river channel. During the dry season, C. orinocensisand C. temensisfed heavily on small characiform fishes, and C. intermedin consumed a variety of fishes, including loricariid catfishes. Only two peacock cichlids, C. orinocensisand C. temensis,were found in the Pasimoni, a more acidic and less seasonal ecosystem. These two species were larger in the Pas- imoni compared with Cinaruco conspecifics; however, conspecifics from the two locations conformed to similar mass-length relationships. In both river systems, C. temensiswere significantly larger than congeners, and C. temensisconsumed larger prey than congeners in the Cinaruco. Examination of gonads and sizes of intra- ovarian oocytes indicated that reproduction by all five populations probably is ini- tiated during the late low water period and may continue into the flooding period.

Investigamos la ecologia de los pavones (Cichlidae, genero Cichla), picivores diur- nos importantes en la pesca deportiva, en el Rio Cinaruco (cuenca del Rio Orinoco en el sur del estado Apure, Venezuela) durante la sequia de los alios 1992-1993, y del Rio Pasimoni (cuenca Rio Amazonas, estado Amazonas) de Venezuela durante los primeros meses del 1993. En el Cinaruco, tres especies de Cichla reparten los habitat disponibles: C. intermedin se ubica cerca de cobertura (arboles sumergidas, etc.) en el cauce principal cerca 0 dentro de la corriente rapida; C orinocensisse encuentra mayormente en agua de poca profundidad cerca de la orilla de lagunas 0 en el cauce principal sin mucha corriente; y C. temensisocurre en agua mas pro- funda allado de bancos de arena y en lagunas y el cauce principal donde la orilla tiene una pendiente alta y es rocosa ("ripio 0 arrecife" de apurelio). Durante la sequia, C. orinocensisy C. temensisse alimentaron principalmente de pequelios peces caracidos, y C. intermedin consumi6 una variedad de peces, incluyendo corronchos loricoridos. Solamente dos pavones C. orinocensisy C. temensis,se encontraron en el rio Pasimoni, un ecosistema con menos fluctuaci6n estacional y mas acido. Estas dos especies alcanzaron tamanos mas grandes en el Pasimoni compararon con los pavones del rio Cinaruco de la misma especie, sin embargo, su curva tallo-peso era la misma. En ambas cuencas, los C. temensiseran significativamente mas grande que los otros pavones. En el rio Cincaruco, los C. temensisconsumieron presas mas gran- des que las de los otros dos pavones. Examinaci6n de la g6nades y los tamanos de los huevos indica que la reproducci6n en las cinco poblaciones de pavones estudia- dos probablemente se inicia durante la parte final de la sequia (aguas bajas) y con- tinua durante la etapa de lluvias e inundacion.

P EACOCK cichlids (Cichla spp., Cichlidae) cies from the Amazon Basin require description are one of the most widespread and abun- (Kullander and Nijssen, 1989). Up to three co- dant predatory fishes in clearwater and black- existing Cichla species have been found in the water ecosystemsof the Orinoco and Amazon upper Rio Orinoco and upper Rio Negro basins Basins. Currently, systematists recognize five in southern Venezuela (Machado-Allison, 1971). Cichla species (known as pavon in Venezuela Peacock cichlids are diurnal piscivores that and'3s tucunare in Brazil), but additional spe- consume a variety of prey and potentially can

f) 1997 by the American &JCiety of Ichthyologist, and HerpetoK'gists WINEMILLER ET AL.-ClCHIA ECOLOGY IN VENEZUELA 691

O"LaJ

Fig. 2. Map showing the locations of the Rio Cin- aruco and Rio Pasimoni in southern Venezuela.

Fig. 1. (A) The banded peacock , Cichla Ie- of these important Neotropical piscivores is merl.5is(81.5 cm SL), from the Rio Pasimoni; (B) royal based more on nonindigenous populations in peacock cichlid, Cichla intennedia (39 cm SL), from reservoirs than on fishes in their native fluvial the Rio Cinaruco; and (C) butterfly peacock cichlid, environments. We examined local population (43 cm SL), from the Rio Pasimoni. structure, body mass-body length relationships, age at length, reproductive condition, patterns of habitat use, and diets of sympatric Cichlaspe- influence prey populations. For example, the cies in two rivers of southern Venezuela. During introduction of exotic peacock cichlids into the dry season, Cichla ecology was investigated Lake Garon, Panama, was followed by massive in a river draining the southern savannas (low changes in community composition and the llanos) and a river that drains forest watersheds structure of the aquatic food web (Zaret and in southernmost Venezuela. Three species (C. Paine, 1973). A similar scenario of community intermedia, C. orinocensis,C. temensis,Fig. 1) are transition followed the introduction of C. ori- sympatric in the Rio Cinaruco (savanna). The nocensisand C. temensisinto Las Majaguas (AB-D, same three species coexist in rivers of forested Universidad Experimental de Los Llanos Occi- regions to the south (Machado-Allison, 1971), dentales thesis, unpubl., 1986; Winemiller, but C. intermediaapparently is absent from our 1989), Guri (Novoa et al., 1989; Gil et al., 1993), Rio Pasimoni study site. and other reservoirs in Venezuela. Zaret (1980) described the life history of a METHODS AND STUDY SITES nonindigenous population of C. ocellarisin Lake Garon, Panama. Characteristics of nonindigen- The Rio Cinaruco in southern Apure State ous C. ocellaris populations in Hawaiian and (R. Orinoco Basin) and the Rio Pasimoni in Puerto Rican reservoirs were summarized in southern Amazonas State of Venezuela (R. technical reports (W. Devick, DLNR re- Casiquiare-R. Negro-R. Amazonas Basin; Fig. 2) port, unpubl., 1972; C. Lilyestrom et al., Puerto are low-gradient, blackwater rivers. The Cina- Rico DNER report, unpubl., 1994), and the life ruco originates in the flatlands of eastern Co- history of Cichla in fish farms was investigated lombia and flows eastward into Venezuela where in Brazil (Fontenele, 1950; Braga, 1952, 1953). it forms the southern boundary of Santos Lu- Very limited information on Cichla ecology in zardo National Park. The Cinaruco lies within a native rivers has been published (Lowe-Mc- region that experiences markedly seasonalrain- Connell, 1969; Goulding, 1980; Winemiller et fall and contains sandy, nutrient-poor soils, sa- al., 1995). Taphorn and Barbarino-Duque vanna vegetation, and acidic streams. In south- (1993) performed a mark/recapture study of ern Amazonas, the landscape is dominated by Cichla populations in rivers and lagoons of San- evergreen tropical forests. The Pasimoni origi- tos Luzardo National Park in Venezuela. nates on the northern slope of the Neblina 'Currently, our understanding of the ecology mountain range and flows northward to the Rio 692 COPEIA, 1997, NO.4

TABLE 1. WATER QUALITY PARAMETERSFOR THE Rio PASIMONI AND Rio CINARUCO (5-12 FEBRUARY1993).

Rio Pasimoni Rio Cinaruco , Channel Lagoon Channel Lagoon pH 4.6 5.1 5.7 5.8 Conductivity(ltmhos/cm) 30 29 35 35 Total alkalinity (mg/l) <0.5 <0.5 <0.5 <0.5 Hardness(mg/l) <0.5 <0.5 6.0 6.0 Chloride (mg/l) 5.6 5.8 1.5 2.8 Color (ftu) 36 39 4 8 Total inorganic nitrogen (ltg/I) 29.3 30.0 7.0 11.2 Total inorganic phosphorus(ltg/I) 3.7 1.8 0.9 2.7 Temperature 26.0 26.0 30.0 31.0

Casiquiare and is surrounded by low, poorly growth checks on saggital otoliths were deter- drained terrain. Annual rainfall is greater, less mined for a few large individuals (C. temensis-- seasonal, and less predictable in this rain forest Pasimoni n = 4, C. temensi.r-Cinarucon = 3; region than in the llanos. We performed in ten- C. orinocensis--Cinaruco n = 2; C. intermedia-- sive surveys of the Cinaruco during 20-24 jan- Cinaruco n = 2). Recent research at the Cina- uary 1992, 7-12 February 1993, and 31 March ruco (D. jepsen, A&M University, unpubl. to 2 April 1993; we surveyed the Pasimoni dur- thesis, 1995) indicates that growth checks are ing 31 january to 5 February 1993. annuli that form during the late dry season The following environmental parameters (March to April) when fish prepare for spawn- were estimated using meters (YS1, Inc.): con- ing. Most Cichla specimens were handled in the ductivity, salinity, temperature, and dissolved ox- field, but a few were preserved in formalin and ygen. Standard titration methods were used to returned to the Universidad Experimental de measure total alkalinity, chloride, total hard- Los Llanos Occidentales or Texas A&M for ness, pH, water color, and total dissolved inor- stomach contents analysis. ganic nitrogen and phosphorus; and visual es- timates were used to record substratum com- position, submerged woody structure, channel RESULTS width, water depth, and relative current velocity. Fishes were sampled in the river channel and Water samples from the Pasimoni and Cina- lagoons with seines, gillnets, castnets, and hook ruco revealed classic blackwater features (Sioli, and line. Angling with lures produced almost all 1984) of low pH, conductivity, solute, and nu- of our Cichla specimens (SL > 20 cm), and al- trient concentrations (Table 1). Water chemis- though no juvenile Cichla were captured, a few try varied little between lentic lagoon and lotic juveniles were observed in shallow backwaters channel habitats within each river. and low velocity shoreline habitats of the Cina- In both river systems, C. temensiswere signifi- ruco channel. Sites were sampled to represent cantly larger than congeners (Fig. 3), and C. or- all major habitats (i.e., river channel, backwa- inocensisand C. intermedia had similar size dis- ters, lagoons, tributary creeks). A single C. ori- tributions in the Cinaruco. Mass-length relation- nocensiswas taken from a Cinaruco creek and is ships of the three species did not differ (AN- not included in this dataset. The remaining CaVA F-tests for regression slopes and sampling effort was divided approximately intercepts with loge-transformed data, n = 84, equally between river channel versus lagoons 77,26, P> 0.05). Even though C. temensisand and backwaters. All captured Cichla were mea- C. orinocensisfrom the Pasimoni had larger av- sured (SL to nearest 0.5 cm) and weighed erage and maximum sizes compared with con- (nearest 5 g). Voucher fish specimens were pre- specifics from the Cinaruco (Fig. 3), conspecif- served in 15% formalin and later deposited in ics from the two rivers conformed to similar the MCNG, Guanare, Venezuela; TCWC, Col- mass-length regressions. The loge-loge regres- lege Station, Texas; UF, Gainesville, ; and sion equations were as follows: C. temensi.r- INHS, Champaign, Illinois (institutional abbre- Pasimoni, M = 3.2L - 4.6, r2 = 0.98; C. temen- viationsfollowLevitonetal., 1985). si.r-Cinaruco, M = 3.2L - 4.7, r2 = 0.96; C. For each Cichla species, a subsample from the orinocensis--Pasimoni, M = 2.9L - 3.4, r2 = total spectrum of size classeswas sacrificed for 0.90; C. orinocensi.r-Cinaruco, M = 2.9L - 3.5, stomacb contents analysis; plus the number of r2 = 0.85. Slopes and intercepts were not sig- WINEMILLER ET AL.-ClCHLA ECOLOGY IN VENEZUELA 693

101 - :C. :;'~i;~temensis - Pasimoni 120 .RIO C.Cinaruco temensis 5 "C 100 . C. orinocensis ~ D C. intermedia .! 80 0 '0 251 _..111":'~. Cmaruro u~ 60 3O

20 ~ 40 15 C. temensis - Cinaruco ~ 10 z 20 5

150 .." . 0 Channel Lagoon >- 30 g 10 RIo Pasimoni ~ C.orinocensis - Pasimoni "C 25. C. temensis ~ 5 ~ . C. orinocensis u. g 20 0 ~ 2520 l._III_Cori~';'- :m:ru: U... 15 15 C. orinocensis - Cinaruco ~ 10 10 ~ Z 5 5 0 15 1 JI. c mt~- - Cm.moo 0 Channel Lagoon 10 C. intermedia - Cinaruco Fig. 4. Numbers of individuals of three Cichlaspe- 5 cies sampled from two habitats of the Rio Cinaruco and Rio Pasimoni. Sampling effort was divided ap- 0 proximately equally between channel and lagoon hab- 2024283236404448525660646872768084 itats. The association between species and habitat was Length Interval (cm) significantfor the Cinaruco (X2= 101.4,df = 2, P < . 3 h ... . 0.0001). Fig. . Lengt frequencydistributions of ClchlaIe- mensisand C. orinocensispopulation samplesfrom the Rio Pasimoni and Rio Cinaruco and C. intennedia from the Cinaruco. Statisticalcomparisons are given Based on Cinaruco data, the minimum standard in the text. lengths of fishes with ripe gonads were as fol- lows: C. temensis,male = 36 cm, female = 32.5 cm; C. orinocensis,male = 31 cm, female = 27.0 nificantly different (ANCOVA F-tests, Range for cm; and C. intermedia male = 32 cm, female = n = 17-60, P> 0.05). 27 cm. Otoliths from C. ternensis(68-75 cm SL, 6.5- Cichla temensisand C. orinocensisoccupied 9 kg) from the Pasimoni had 7-9 annuli. If one both channel and lagoon habitats in the Pasi- assumesthat maturation occurs at 1-2 yr (Zaret, moni and Cinaruco (Fig. 4). During our dry 1980), then our Pasimoni C. temensissample was season surveys, C. orinocensisusually was ob- dominated by 9-11-yr fish that grew at an aver- served and captured in shallow nearshore areas age of about 7 cm and 0.75 kg per yr. of lagoons and slow moving reaches of the river In our january/February Cinaruco samples, channel. Cichla temensistended to occupy deep- most individuals of all three specieshad inactive er littoral areas in lagoons and along sandy and gonads or gonads in intermediate stages of re- rocky banks of the main river channel. In the crudescence, and very few had fully ripe gonads Cinaruco, C. intermediawas alwayscaptured near (no C. orinocensiswere ripe). In our March/ submerged rocks or woody structure in the pri- April 1993 samples, 11 of 27 C. temensiswere mary river channel within or near swift current. ripe, three of 13 C. orinocensiswere ripe, and Mean lengths of C. temensisand C. orinocensisdid four of 10 C. intermediawere ripe. During the not differ between habitats (Fig. 5); however C. january 1992 and April 1993 surveys of the Cin- temensiswas significantly larger than C. orinocen- aruco, one or more individuals of both C. inter- sis in lagoons (t = 5.21, df = 23, P < 0.0001) media and C. temensiswere observed attending and the river channel (t = 2.14, df = 13, P = nests with eggs. On 2 February 1993, we ob- 0.05) and C. intermediain the river channel (t = serveda single C. orinocensison a nest with eggs 2.91, df = 32, P < 0.01). ne~ the shoreline of a lagoon of the Pasimoni. Small fishes «: 10 cm SL), especially the 694 COPEIA, 1997, NO.4

~E 50 . c. te~ensis . mensis as they attacked a shoal of S. kneri that !:- C. onnocensis.. had exited.. a lagoon Into a broad, shallow reach "" 45 of the Cinaruco. On a few occasions at the Cin- ~ aruco, we observed C. orinocensis beach them- ~ 40 selvesduring prey pursuit. "E 35 The prey fields of the two river systemswere ~ similar, but fish densitieswere lower in the Pas- ~ 30 imoni. Five sites in the Pasimoni yielded 80 fish IJ) species from 23 families; 31 sites in the Cina- m 25 ruco yielded 211 fish speciesbelonging to 35 ~ families (voucher material and data in MCNG, 20 Channel Lagoon INHS), and only about 40 species occurred in both rivers. The Pasimoni has not been sur- .Fig. 5. ~ean sta~dar.d lengt~s (~th SD bars) of veyed as extensively as the Cinaruco, and addi- Czchlaspecies ~y habitat m ~e Ri<:>Cmaruco. (Sampl.e tional species probably inhabit the former. We n for C. temenszs= 98, C. onnoCe1lSzs= 62, C. mtermedza b d h. h fi h d. ks. th . - 35) S . . I I .. th 0 serve Ig er s stan Ing stoc In e ClP- - . tatlstlca resu ts are given m e text. d . h th P . . D . aruco compare WIt e aSlmOnl. unng the dry season, a single 10 m seine haul (6.1 X 1.8 characids, formed the major diet component of m net) made either during day or night in the all three Cichla species (Table 2). Cichla inter- Cinaruco typically yielded from 15-30 species media consumed a greater fraction of catfishes, and 50-500+ individuals. In shallow shoreline primarily small loricariids, than did the other areas of the Cinaruco, hundreds of individuals two species (Table 2). The species most fre- and dozens of specieswere viewed at night with quently found in C. temensisstomachs was Hem- the aid of a flashlight. Far fewer individuals and igrammusanalis (Characidae, n = 5), whereas H. slightly fewer specieswere observed in the same vorderwinkleriwas most frequent in C. orinocensis manner within similar shoreline habitats of the stomachs (n = 7). All three Cichla species con- Pasimoni. Characiformes and Siluriformes dom- sumed H. analis and small Crenicichlasp. (Cich- inated both river assemblages,comprising 52- lidae, probably C. notopthalmus). Nannostomus 60% and 16-20% of species, respectively. Cich- sp. (Lebiasinidae, n = 4) and anostomids (n = lids also were an important group (17% of spe- 4) were found in stomachs of C. temensisand C. cies in the Pasimoni; 10% in the Cinaruco). orinocensis.Average prey length was greatest for C. temensis,followed by C. intermedia (Table 2). DISCUSSION Cichla temensisprey were significantly larger than C. orinocensisprey (t = 2.10,df = 25, P < 0.05). Although both would be characterizedas On several occasions, we observed large (60- blackwater ecosystems (low pH, humic acid 80 cm SL) C. temensisattacking large characi- stained, low conductivity; Sioli, 1984), the two riv- forms (e.g., Semaprochiladuskneri [Prochilodon- ers have very different watersheds.The Cinaruco tidae] and Acestrorhynchusheterolepis [Characi- drains savanna habitats where precipitation is dae] approximately 20-35 cm SL) in shallow strongly seasonal (virtually all rain occurs from shoreline areas. We captured several large C. te- May to August). The Cinaruco has a mostly ev-

TABLE 2. PERCENT NUMBER OF PREY CATEGORIESAND MEAN TOTAL LENGTHS OF PREY IN Cichla DIETS DURING THE DRY SEASONAT THE Rio CINARUCO. Values inside parentheses are percentages based on total numbers of prey excluding unidentified fishes.

c. /emtn5U c. urinocenJis C. intennedia

Characiforms 53.8 (73.7) 81.5 (88.0) 61.5 (72.7) Siluriforms 3.8 (5.3) 0 (0) 15.4 (18.2) Perciforms 15.4 (21.0) 11.1 (12.0) 7.7 (9.1) Unidentified fishes 26.9 7.4 15.4 Mean prey TL (cm) 4.62 3.34 3.72 Standard deviation 2.59 2.15 1.81 Number stomachs examined 87 51 31 Number empty stomachs 66 32 23 Total number prey items 26 27 13 , WINEMILLER ET AL.-ClCHLA ECOLOGY IN VENEZUELA 695 ergreen gallery forest. The Pasimoni drains a ly, mortality rates may be similar, but C. temensis heavily forested region that receives less seasonal might grow faster than C. orinocensisand C. in- precipitation (driest months are December to termedia.At annulus 5, Cinaruco C. temensiswere February). When compared with Cinaruco con- larger (mean = 54 cm SL and 3.5 kg) than C. specifics, a greater proportion of Pasimoni C. te- orinocensis(mean = 35 cm SL and 0.975 kg) and mensiswere captured from lagoons. Precipitation C. intermedia (mean = 39 cm SL and 1.15 kg), occurs frequently during the dry seasonin south- which is consistent with the latter mechanism. ern Amazonas, and the water velocity in the main The primary nesting habitat appeared to be channel of the Pasimoni may exceed the prefer- the littoral regions of lagoons (C. temensisand C. enda of Cichlaor their prey. Additionally, the pro- orinocensis)and slow moving side channels and ductivity of lagoons may be higher than the river back eddies along the shore of the main channel channel to a greater extent in the Pasimoni than (C. temensis).When not nesting or brooding, the Cinaruco. Although prey densities seemed adult Cichla partitioned habitat during the dry higher in the Cinaruco, this apparent difference season, with C. intermediarestricted to flowing was not reflected in between-site comparisons of channel habitats with submerged structure, C. or- body condition. inocensisin lagoons and shallow lentic areas of Although Cichla body mass-body length rela- the river channel, and C. temensisin deeper lit- tionships were virtually identical for conspecifics toral regions of both the channel and lagoons at the two locations, Pasimoni Cichla were con- (see also Taphorn and Barbarino-Duque, 1993). siderably larger than their Cinaruco conspecif- Although large prey may be encountered occa- ics. This pattern was particularly strong for C. sionally in the shallow-waterhabitats occupied by temensis.Only one C. temensissmaller than 60 cm C. orinocensisand C. intermedia,there would be SL was captured at the Pasimoni, whereas only many more opportunities to encounter larger one C. temensislarger than 60 cm SL was cap- prey, especially shoals of migrating Semaprochila- tured from the Cinaruco. A few C. temensisex- dus, in the deeper littoral areas occupied by C. ceeding 68 cm SL and 8 kg have been taken by temensis.Larger mean prey size at the sameintake anglers at the Cinaruco. Cichla orinocensishas rate and energy density would represent greater the potential to attain a greater maximum size total energy intake. The high percentage (66- than those indicated by data from the Cinaruco 75%) of empty stomachsdocumented during the and Pasimoni populations, as evidenced by a late dry season at the Cinaruco indicates that 55-cm SL, 5-kg specimen captured by one of us prey capture may be infrequent for all three spe- (KOW) from Las Majaguas reservoir in western cies during that time of year. Venezuela. Because of the remoteness of the re- Cichla temensisand C. orinocensisin our Cina- gion and government regulations restricting ac- ruco sample consumed more characids and few- tivities there, most Cichla populations in south- er cichlids than those studied by Novoa et al. ern Venezuela had received little impact from (1989) and Gil et al. (1993) in Guri reservoir. fishing until very recently. Therefore, it seems Although characids remain in Guri, their spe- likely that mortality for the largest size classes cies richness and population abundances have of Cichlais greater in the Cinaruco than the Pas- declined relative to cichlids (Alvarez et al., imoni. 1986), as might be expected given the basic life- The among-species differences in the size history differences of these two groups (Wine- structure of Cinaruco Cichla populations could miller, 1989). Cichlids are brood guarders; and, be influenced by ecological differences. The in the absence of strong environmental season- largest species, C. temensis,occupied the greatest ality, some species nest at frequent intervals. range of habitats and consumed the largest Our Cinaruco data for gonad condition indi- prey, including the migratory prochilodontid, S. cate that a portion of each Cichla population kneri. Because mouth gape limits maximum prey was preparing to spawn near the end of the dry size in these whole-fish swallowers, C. temensis season (April). During April 1993, we observed may be the only Cichla that regularly surpasses a few schools of larval C. temensisaccompanied a size threshold that allows it to exploit these by adult pairs. Yet, not all adults were in the seasonally abundant, large prey. same state of gonadal recrudescence during the Two alternative mechanisms could account late dry season, so some individuals may fail to for the observation that C. temensisgrows larger reproduce, or they spawn during a different pe- than its two congeners. Growth rates could be nod of a given year. the same for all three species, but the largest C. Densities of most prey taxa probably peak as temensismay be older than the largest C. orino- waters recede from the broad floodplain of the censisand C. intermedia (meaning that the for- Cinaruco (November to January). Those Cichla ~er's adult mortality rate is lower). Alternative- that fail to optimize their intake of abundant 696 COPEIA, 1997, NO.4

prey during this falling water period could have LEVITON,A. E., R. H. GIBBSJR., E. HEAL,AND C. DAW- difficulty capturing enough energy during the SON.1985. Standardsin herpetologyand ichthy- ensuing dry season to meet the requirements ~Iogy: Part I. Standar~s~bolic codesfor ins~tu- for successful reproduction. During the falling tlonal resource.collections In herpetologyand Ich- water period, immigrating Semaprochilodusmay thyology.Copela 1985:802-832. .. .d U. I I . f LoWE-McCONNELL, R. H. 1969. The clchhd fishes of

proVI e a par cu ar y Important source 0 nu- G So th Am . .th th ...... uyana, u enca, WI notes on elr eco I - tnuon for C. temensasIn these oligotrophic nv- ogy and breeding behavior. Zool. J. Linn. Soc. 48: ers. In many regions of the Amazon and Ori- 255-302. noco basins, these abundant detritivores under- MACHADO-ALLISON,A. 1971. Contribucion aI conoci- go annual migrations from their spawning and miento de la taxonomia del genero Cichla(Perci- nursery grounds in productive floodplains of formes:Cichlidae) en Venezuela.Parte I. Acta Biol- mainstem whitewater rivers into nutrient-poor, ogica,Venezuela 7:459-497. blackwater tributaries (Goulding et al., 1988; NOVOA,D. ~.'J KOONCE,A. LOCCI,~D F. ~OS. Vazzoler et al., 1989; Ribeiro and Petrere, 1990). 1989. La ~ctlofauna~ellago de Gun: comp~slclon, abundanclay potencialpesquero. II. Evaluaclondel potencial pesquerodel lage Guri y estrategiasde ACKNOWLEDGMENTS ordenamientopesquero. Sociedad de CienciasNa- . .. turales La Salle 49:159-197. T.~e Nau?nal Geographical Society, La Fun- RIBEIRO,M. C. L. DEB., ANDM. PETREREJR. 1990. dacI<:>nflUVIal de Los ~anos, and ASOMUSEO Fisheriesecology and managementof the jaraqui pr~VI~ed .support fo~ thiS research. PROFAUNA (Semaprochilodustaeniurus, S. insignis) in central (Mlnlsteno de Amblente y Recursos Naturales Amazonia.Regulated Rivers: Res. & Man. 5:195-215. of Venezuela) provided scientific collection per- SIOU,H. 1984. The Amazonand its main affluents: mits to the authors. Several individuals assisted hydrography,morphology of the river courses,and with fieldwork, especially P. Rodriguez andJ. Ly- river types.p. 127-166.In: The Amazon,Monogra- ons. We thank G. Webb, C. Lofgren, and the phiae Biologicae.Vol. 56. Sioli, H. (ed.). Dr. W. Cinaruco Fishing Club for rendering assistance Junk, The Hague,The Netherlands. d . fi ld t d. N . al TAPHORN, D. C., AND A. BARBARINO-DuQUE. 1993. Ev- unng our e s u les. utnent an yseswere I . , d I . . , I d I . £ d b D Sh W th k a uaclon e a sltuaclon actua e os pavones (Clch- per orme y . ormann. e an D. Jep- la spp.), en el Parque Nacional Capanaparo-Cina- sen, L. Nico, and L. Page for reviewing and ruco, estadoApure, Venezuela.Natura 96:10-25. commenting on an earlier draft. VAZZOLER,A E. DEM., S. A AMADIo,AND M. C. DAR- ACIOLo-MALTA.1989. Aspectos biologicos de peix- LiTERATURE CITED es Amazonicos. XI. Reprodu~ao das especiesdo ge- nero Semaprochilodus(Characiformes, Prochilodon- ALVAREZ,E., L. BALBAS,I. MASSA,AND]. PACHECO. tidae) no baixo Rio Negro, Amazonas, Brasil. Rev. 1986. Aspectos ecologicos del embalse Guri. Inter- Brasiliera Bioi. 49:165-173. ciencia 11:325-333. WINEMILLER,K. O. 1989. Patterns of variation in life BRAGA,R. 1952. Ninhos de tucunares Cichla temensis history among South American fishes in seasonal Humboldt e Cichla ocellarisBloch & Schneider. Rev. environments. Oecologia 81:225-241. Brasiliera Bioi. 12:273-278. -, L. C. KELSO-WINEMILLER,AND A. L. BRENKERT. -. 1953. Crecimiento de tucunare pinima, Cich- 1995. Ecological and morphological diversification la temensisHumboldt, em cativiero (, in fluvial cichlid fishes. Environ. Bioi. Fish. 44:235- Cichlidae). Dusenia 4:41-47. 261. FONTENELE,O. 1950. Contribucao para 0 conheci- ZARET,T. M. 1980. Life history and growth of Cichla mien to de biologia dos tucunares, Actinopterygii ocellaris,a predatory South American cichlid. Bio- Cichlidae, em captiviero. Aprelhlo de reproducao. tropica 12:144-157. Habitos de desova e incubacao. Rev. Brasiliera Bioi. -, AND R. T. PAiNE. 1973. Species introduction 10:503-519. in a tropical lake. Science 182:449-455. GIL, C. E.,]. ANDRADA, E. MENDEZ, AND]. M. SALAZAR. 1993. Estudio preliminar sobre alimentacion en (KOW) DEPARTMENT OF WILDLIFE AND FISHER- cautiverio y contenido estomacal de Cichla temensis IES SCIENCES,TEXAS A&M UNIVERSITY, COL- del embalse Guri, Estado Bolivar, Venezuela. Natu- LEGE STATION, TEXAS 77843; AND (DCT, ra 96:42-47. AB-D) PROGRAMAPOSTGRADUADO EN MANE]O GoUI:DING:M. 1980. .The fishes ~d the for~st: explo- DE RECURSOSNATURALES RENOVABLES,UNIV- rations In Amazoman natural history. Umv. of Cal- Ilorma.J: . P ress, L os An geI es. ERSIDAD E XPERiMENTAL DE LOS L LANOS 0 C- -, M. L. CARVALHO,AND E. G. FERREIRA.1988. CIDENTALES ESEQUIEL ZAMORA, GUANARE, Rio Negro: rich life in poor water. SPB Academic, PORTUGUESA, VENEZUELA 3310. E-mail: The Hague, The Netherlands. (KOW) [email protected]. Send re- KULLANDER,S., AND H. NI]SSEN.1989. The Cichlids print requests to KOW. Submitted: 12 June of Surinam: Teleostei, Labroidei. E.]. Brill, Leiden, 1996. Accepted: 28 April 1997. Section edi- The ~etherlands. tor: S. T. Ross.